Declared “the top word or phrase of the year 2023 thus far” by The Global Language Monitor (2023), artificial intelligence (AI) is one of the most common tech buzzwords of our time. However, its popularity in common parlance raises a significant question: how far have we, as a country, reckoned with the societal and ethical implications of this burgeoning technology? What happens when AI transcends the limits of technological advancement and enters the realm of malicious intent and cybercrime? Are we equipped with adequate legal and digital safeguards to protect ourselves from its potential adverse effects?
A retired central government official from Kozhikode, Kerala, receives an unknown WhatsApp call. Seeing as the caller is anonymous, he is initially dismissive and refuses to pick up, but is soon surprised to find a slew of messages from the same contact, now disclosing his identity as a former colleague. They exchange text messages for a while, and the supposed colleague soon dials another voice call. The official, who was feeling wary of a spam or prank call until now, is relieved to find that the person on the other end does sound like his former colleague. His suspicions are renewed soon, however, when the colleague begins to request a financial favour on the basis of an elaborate story. To placate his fears completely, the two get on a video call, and the official is assured that the person he is talking to is in fact his old friend, and with such assurance, he completes a transaction of ₹40,000 into the latter’s bank account.
As it would turn out, the official’s initial fears were not misplaced, and the caller at the other end had been a young Gujarati man who was able to defraud the former using highly sophisticated, AI-powered technology called ‘deepfake’.
A portmanteau of ‘deep learning’ and ‘fake’, deepfakes refer to an AI-generated form of synthetic media where a person’s likeness in audio or video can be replaced by another using advanced machine learning models that manipulate large sets of audiovisual data. While fabricated photos and videos are not a new phenomenon on social media, the striking accuracy and feigned authenticity that deepfakes are able to achieve can fool even the most well-versed of tech users. Deep Fakes are able to do so by relying on neural networks and Generative Adversarial Networks (GANs) that map content from one person to another by finding common features. While the technology behind it is evidently complex, deep fakes are becoming increasingly accessible due to the rise of open access software and apps such as Lensa AI, DeepFaceLab, and Reface, among others.
Although the incident with the cunning scammer mentioned at the beginning of this article is unsettling enough to keep the common man awake, the depths of deception and ramifications of deepfake technology go far beyond what most can envision. Right at the outset, concerns over the malicious use of deepfakes first arose with the rise of deepfake pornography on Google and Microsoft’s search engines. When the Italy-based AI firm Deeptrace Labs conducted a study on deepfake media that was proliferating on the internet, it found that 96% of all deepfake videos were pornographic in nature, with almost 100% of those involving women whose likeness had been morphed onto sexually explicit videos without their consent. This statistic has alarming implications for cybercrimes against women, where such videos can be used for harassment, coercion, and public humiliation. The most recent and viral deepfake video of actress Rashmika Mandanna, which garnered significant outrage from Indian netizens, indeed epitomises the intersectional threat of deepfakes as a technology for online gender-based violence.
Furthermore, deepfakes hold the potential to constitute a very real and plausible political threat. Social media platforms are replete with manipulated videos of politicians apparently uttering words that they never said. While a large majority of these are currently inoffensive and easily distinguishable as fake, the underlying potential for harm is indisputable. As a political tool for disinformation, deepfakes can be used to discredit notable figures, mislead the public, and erode their trust in online media sources.
The political threat of deepfakes is not limited to domestic politics, but extends to the international stage, where the dissemination of misleading and manipulated information can have a profound impact on sensitive geopolitical situations. Citing deepfake videos of the Ukrainian President Zelenskyy and Russian President Putin that were going viral at the height of the Russia-Ukraine conflict, the Delhi-based think tank, Observer Research Foundation (2023), concluded that the use of generative AI can have “grave implications for national security and law enforcement” when utilised in situations of conflict where credible information is further limited. In the worst-case scenario, it can be used to instigate or escalate crisis on-ground by exploiting digital vulnerabilities.
In the light of such emerging threats, it will be crucial to inculcate a new form of online media literacy to empower a vigilant citizenry. Currently, the misuse of deepfake technology is only limitedly covered by the cyber protection laws of India, including Section 66E and 66D of the IT Act, and certain sections of the Indian Copyright Act. These primarily pertain to the infringement of one’s privacy and unauthorised use of their personal property. The way forward must involve the development of more robust legal frameworks that can specifically target and mitigate this rising crime of the future.
Sources
Jaiman, A. (2023, January 1). The danger of deepfakes. The Hindu. Retrieved From https://www.thehindu.com/sci-tech/technology/the-danger-of-deepfakes/article66327991.ece
Johnson, D., & Johnson, A. (2023, June 23). What are deepfakes? How fake AI-powered audio and video warps our perception of reality. Business Insider. Retrieved From https://www.businessinsider.com/guides/tech/what-is-deepfake?IR=T
Khandekar, O. (2021, January 1). How deepfake porn and propaganda threatens India. Live Mint. Retrieved From https://lifestyle.livemint.com/news/big-story/deepfakes-when-seeing-is-not-believing-111609504596030.html
Patil, S., & Gori, S. (2023, August 25). Deep fake, disinformation, and deception. Observer Research Foundation. Retrieved From https://www.orfonline.org/expert-speak/deep-fake-disinformation-and-deception/
Romano, A. (2019, October 7). Deepfakes are a real political threat. For now, though, they’re mainly used to degrade women. Vox. Retrieved From https://www.vox.com/2019/10/7/20902215/deepfakes-usage-youtube-2019-deeptrace-research-report
- Sharma, V. (2023, August 23). Kerala deepfake fraud case: Efforts on to nab accused from Gujarat, say cops. Hindustan Times. Retrieved From https://www.hindustantimes.com/india-news/aibased-deepfake-scammer-identified-accused-of-cheating-elderly-man-in-kerala-police-launch-manhunt-101692036702353.html
The web is drowning in Deepfakes and almost all of them are porn. (2019, October 13). Wired. Retrieved From https://www.wired.co.uk/article/deepfakes-porn
Comments
Post a Comment